ECOLOGICAL CONSTRAINTS ON CIllNSTRAP PENGUIN FORAGING BEHAVIOR: THE ROLE OF DIEL AND SEASONAL LIGHT CHANGES by JOHN KEVIN JANSEN A THESIS Presented to the Department ofBiology and the Graduate School of the University ofOregon in partial fulfillment of the requirements for the degree of Master of Science March 1996 ; "Ecological Constraints on Chinstrap Penguin Foraging Behavior: the Role of DieI and Seasonal Light Changes," a thesis prepared by John Kevin Jansen in partial fulfillment of the requirements for the Master of Science degree in the Department of Biology. This thesis has been approved and accepted by : ~CHc~~~A& Dr. Nora TerwIllIger, Chair of Exammmg Comnuttee 27 February 1996 Date 11 Committee in charge: Accepted by: Dr. Nora Terwilliger, Chair Dr. Janet Hodder Dr. Lynda Shapiro Dr. Peter Boveng Vice Provost and Dean of the Graduate School © 1996 John Kevin Jansen III IV Title: ECOLOGICAL CONSTRAINTS ON CHINSTRAP PENGUIN FORAGING BEHAVIOR: THE ROLE OF DIEL AND SEASONAL LIGHT CHANGES March 1996 Master of Science to be taken for the degree of An Abstract of the Thesis of Approved: ~(Hr~T~A& Dr. Nora Terwilliger John Kevin Jansen in the Department of Biology The relationship between foraging performance and prey availability can only be accurately interpreted by understanding the environmental constraints on predator behavior. Penguins are considered visual predators, but little is known of how their foraging tactics are influenced by changes in light. Chinstrap Penguin foraging was examined during two breeding seasons to test the hypothesis that light is a constraining factor. Radio telemetry of penguin arrivals and departures revealed diurnal and overnight foraging. Overnight foraging was common during chick-brooding, but after chicks creched, penguins foraged mostly during the day suggesting a preference for diurnal feeding. Diet sampling indicated that birds were capturing mostly krill, but that fish were more common in overnight foragers. Penguins foraging by day were found to spend more time at sea during reduced light levels in one season, but not the other. This study shows that light plays a major role in determining the daily foraging of Chinstrap Penguins and the availability of its prey, but that its effects are often mediated by other constraints. CURRICULUM VITA NAME OF AUTHOR: John Kevin Jansen GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon University ofWashington University of Colorado DEGREES AWARDED: Master of Science in Biology, 1996, University of Oregon Bachelor of Arts in Population and Qrganismal Biology, 1987, University of Colorado SPECIAL AREAS OF INTEREST: Foraging and Reproductive Ecology ofMarine Birds PROFESSIONAL EXPERIENCE: Teaching Assistant, Oregon Institute ofMarine Biology, University of Oregon, Charleston, 1994-95 Wildlife Biologist, National Marine Mammal Laboratory, Seattle, Washington, 1990-present v Fisheries Technician, Washington State Department ofFisheries, Seattle, Washington, 1988 Biological Technician, Loxahatchee Wildlife Refuge, Loxahatchee, Florida, 1988 v VI AWARDS AND HONORS: Marcia Brady Tucker Award, American Ornithological Union, 1995 Outstanding P~rformanceAward, National Marine Mammal Laboratory, 1991-92, 1995 GRANTS: Graduate Teaching Fellowship, Summer Session, University ofOregon, 1994-95 PUBLICATIONS: Bengtson, J. L., P. L. Boveng, and J. K. Jansen. 1991. Foraging areas ofkrill- consuming penguins and fur seals near Seal Island, Antarctica. Antarctic Journal 1991 Review:217-218. Croll, D. A~ Hewitt, R. P.~ Derner, D. A, and Jansen, J. K. 1993. Penguin foraging behavior in relation to the distribution of prey. Report of the Working Group for the CCAMLR Ecosystem Monitoring Program~ Seoul, Korea. CCAMLR. Croll, D. A, J. K. Jansen, and 1. L. Bengtson. 1991. Reproductive performance of Chinstrap Penguins on Seal Island, South Shetland Island, Antarctica. Antarctic Journal 1991 Review:218-220. Croll, D. A, J. K. Jansen, 1. L. Bengtson, P. L. Boveng, and M. G. Goebel. 1996. Foraging behavior and reproductive success in Chinstrap Penguins: the effects of transmitter attachment. Journal ofField Ornithology 67(1):in press. Meyer, W. R., J. L. Bengtson, 1. K. Jansen, and R. W. Russell. 1996. Foraging trips and provisioning ability of Chinstrap Penguins: effects ofbrood size and interannual vanability. Polar Biology:in press. Vll ACKNOWLEDGEMENTS This thesis represents the merging of ideas and expertise from many colleagues, both in the field and during various phases of analysis and writing. Many insights arose from discussions with members of earlier field teams, including Don Croll, Mike Goebel, and Steve Osmek, often while observing our ever-fascinating subjects. John Bengtson and Peter Boveng shared a wealth of experience involving field research without which this project would not have been possible. I also thank William Meyer, Mike Schwartz, and Brian Walker for their valuable and enthusiastic contribution during phases of the research at Seal Island. I greatly appreciate the friendship and intellect of all those I lived and worked with at Seal Island who fostered a camp spirit blending immense curiosity about the natural world with an ability to find humor while exploring it. I am grateful for the' ideas generated during discussions with Jan Hodder. Her editorial suggestions greatly improved this manuscript. I also thank Lynda Shapiro and Nora Terwilliger who provided helpful comments during various phases ofwriting and analysis. Peter Boveng contributed valuable insights during data analysis and final editing. I appreciate the analytical expertise ofJean Adamson and Richard Emlet during my surprise office visits. I am also grateful for the encouragement ofmy life's companion, Karen Dumler, who inspired me to challenge myself This research was supported by NOAA's National Marine Fisheries Service, as part of it's Antarctic Marine Living Resource program. I. INTRODUCTION............................................... 1 II. MATERIALS AND METHODS 6 III. RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 18 Page TABLE OF CONTENTS Vlll Characteristics ofForaging Trips 18 Daily Foraging Patterns 24 Individual Foraging Behavior 28 Pattern ofArrivals and Departures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Arrivals and Departures in 1993 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Diurnal Trips . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . 31 Standard Overnight Trips. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Extended Overnight Trips 35 Arrivals and Departures in 1994 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Diurnal Foraging Trips 36 Standard Overnight Trips 36 Extended Overnight Trips 37 Duration ofForaging I:rips 37 Effect ofDeparture Time on Duration ofForaging Trips 38 Light Regimes for Foraging Penguins 41 Effects of Variable Illumination on Duration ofForaging 43 Daily and Seasonal Occurrence ofNocturnal Foraging 43 Diet ofDiurnal and Overnight Foragers 45 Species Description 6 Description of Study Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Measurement ofForaging Patterns 7 Criteria to Exclude Non-Provisioning Foragers 11 Determination ofFledging , ' 12 Food Load Sizes and Diet Composition 13 Estimates ofLight Intensity during Foraging 15 Chapter IV. DISCUSSION 51 BIBLIOGRAPHY 83 Chapter IX Page Critical Assumptions 51 Sampling Regime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Potential Effects ofInstrument Attachment 52 Arrivals and Departures at Night 53 Light-Dependent Foraging 57 Timing and Frequency of Foraging 57 Comparisons with Other Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Seasonal Changes in Foraging Patterns 62 Determinants of Time Spent Foraging 66 Consequence ofActivity Patterns and Departure Time . . . . . . . . . . . . 67 Consequence ofVariation in Available Light 70 Light-Dependent Differences in Diet 74 Summary of Conclusions 79 x3. Mean Foraging Trip Durations for Chinstrap Penguins for All Breeding Periods Combined ~ . . . . . . . . . . . . . . . . . . . . 38 2, Mean Percent Trip Frequency for Chinstrap Penguins Taking Diurnal, Standard Overnight, and Extended Overnight Trips 28 Page LIST OF TABLES 4. Linear Regression Statistics for the Relationship between Foraging Trip Duration and Departure Time for Diurnal, Standard Overnight, and Extended Overnight Trips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 1. Percent Frequency of Occurrence ofDaily Foraging Activity Patterns for Chinstrap Penguins . . . . . . . . . . . . . . . . . 25 5. Diet Mass and Composition and Frequency of Occurrence ofFish in the Diet of Chinstrap Penguins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Table Xl 9. Relationship Between For-aging Trip Duration and Time ofDeparture during 1993 and 1994 , , .. , 39 1. Location of Seal Island within the Antarctic Peninsula Region and the Relative Positions of Study Colonies on Seal Island ................, 8 Page LIST OF FIGURES 8. Distribution ofDeparture and Arrival Times of Chinstrap Penguins'for Diurnal, Standard Overnight, and Extended Overnight Foraging Trips in 1994 , , 33 7. Distribution ofDeparture and Arrival Times ofChinstrap Penguins for Diurnal, Standard Overnight, and Extended Overnight Foraging Trips in 1993 , . , , , , , , 32 5. Incidence ofDiurnal, Standard Overnight, and Extended Overnight Foraging in Individual Penguin's Foraging Behavior in 1993 29 3. Midpoint and Span ofEach Foraging Trip on the Scale of a Single Foraging Cycle versus the Midpoint on the Scale ofDays in 1993 . , . , , 20 4. Midpoint and Span ofEach Foraging Trip on the Scale ofa Singte Foraging Cycle versus the Midpoint on the Scale ofDays in 1994 , 22 6. Incidence ofDiurnal, Standard Overnight, and Extended Overnight Foraging in Individual Penguin's Foraging Behavior in 1994 30 2, Frequency Distribution of Chinstrap Penguins Departing from and Arriving at North Cove Colony in Relation to Time ofDay , . ,. 19 10. Relationship Between Mean Light Intensity and Foraging Trip Duration during 1993 and 1994 , , 42 11 ..Relationship Between Mean Duration ofFirst Diurnal Foraging Trips and Brightness Index for 1993 and 1994 . , , , , . . 44 Figure 14. Relationship Between Daily Frequency ofExtended Overnight Foraging Trips and the Daily Mean Duration ofFirst Diurnal Foraging Trips for 1993 and 1994 48 13. Daily Frequency of Overnight Foraging Trips during Chick-Brooding, Transition, and Creche Periods Relative to the Progression of Creching and the Change in Daylength in 1994. . 47 12. Daily Frequency of Overnight Foraging Trips during Chick-Brooding, Transition, and Creche Periods Relative to the Progression of Creching and the Change in Daylength in 1993. 46 XlI PageFigure 1CHAPTER I INTRODUCTION The life history of penguins is constrained by interactions between the species' biology and changing physical and biological factors of its environment. Fluctuations in the demography and breeding success of penguins are associated with large-scale changes in abiotic factors, such as the Southern Oscillation (Croxall 1992). In addition, the capacity of penguins to forage successfully and provision chicks is linked to the status of prey populations near breeding colonies (Croxall et al. 1988a, Trivelpiece et al. 1990). Studies attempting to relate foraging behavior to environmental factors, however, have typically integrated large portions of the breeding cycle and are therefore not sensitive to changes on shorter time scales (CCAMLR, 1988). Few studies have addressed changes in penguin foraging behavior that might occur within a breeding season (Lishman 1985a, Williams and Rothery 1990, Chappell et al. 1993a, Le Maho et al. 1993) and fewer have attempted to relate these changes to measures of the penguins' physical habitat (Wilson et al. 1993). Examining the nature of, and relationships between, physical and biological elements on short time scales may lead to a more accurate interpretation of larger scale processes. Daily and seasonal changes in light intensity, although likely to be important . environmental determinants of foraging behavior in pengliins (Wilson et al. 1989a), have received little attention. 2In attempts to understand the relationship between foraging behavior and prey availability, the lack ofinformation on at-sea foraging tactics has given rise to a necessary assumption that prey within a penguin's horizontal and vertical foraging range represents available food (Croxall et al. 1988b, Hunt et al. 1992, Croll et al. 1993). Knowledge of the physiological adaptations that facilitate foraging, such as diving capacity, has led to a better understanding ofhow penguins are able to narrow the spatial gap between themselves and their prey (Kooyman et al. 1982, Kooyman and Davis 1987, Kooyman et al. 1992, Chappell et al. 1993a, 1993b), thus limiting the scope of this assumption. Few studies have addressed whether constraints on senses, such as vision, might mediate the detection and capture of prey (Wilson et al. 1989a; 1993). It is commonly assumed, however, that penguins are primarily visual hunters even though absolute visual sensitivity and the importance ofvision relative to other senses has not been established (Martin and Young 1984, Howland and Sivak 1984, Bowmaker and Martin 1985, Wilson et al. 1993). Equally important in relating predator to prey is an understanding of the temporal constraints operating on one or both groups, especially ifan objective is to know when predator and prey overlap. Penguins show physiological and behavioral rhythms related to light-dark cycles. For instance, rhythmic secretion of melatonin by the pineal gland in Ad6lie Penguins (£ygoscelis adeliij,e) is linked to an endogenous clock; a clock that has a 24 hour cycle when birds are exposed to constant darkness, a 12 hour cycle in birds exposed to a 12 hour light: 12 hour dark regime, and is suppressed in birds experiencing continuous daylight (Cockrem 1990). Body temperature, thought to be mediated by melatonin, follows the same pattern. Ad6lie Penguin behaviors within the colony, such as 3restlessness during incubation and ecstatic displays, also appear to have endogenous rhythms related to the diellight cycle (Muller-Schwarze 1968, Derksen 1977). Daily cycles in light intensity thus have potential to constrain the physiology of a penguin in two ways: 1) by altering a bird's ability to use vision in perceiving its surroundings, and 2) by entraining a bird's activity pattern through honnonal control. Given these mechanisms it is not surprising that foraging activity has been linked to diel periodicity oflight levels (Wilson et al. 1989a, Wilson et al. 1989b, Golombek et al. 1991), even at latitudes where colonies are exposed to continuous daylight during the austral summer: Yeates (1971) reported that at Cape Ross (77°S) in January Ad6lie Penguins tend to be away from the colony during the middle of the day. Williams and Rothery (1990) showed that during chick-brooding 96% of Gentoo Penguins (Eygoscelis ~) at Bird Island (54 °S) forage exclusively by day. Other studies on Gentoo and Chinstrap Penguins (pygoscelis antarctica}at King George Island (62°S), Jackass Penguins (Spheniscus demersus) at Saldanha Bay, South Africa (33 °S), and Magellanic Penguins CS... magellanicus) in Argentina (42°S) revealed a similar predominance of diurnal foraging (Trivelpiece et al. 1986, Wilson 1985, Scolaro and Suburo 1994). Further studies on Ad6lie, Chinstrap, and Gentoo Penguins show that birds taking overnight foraging trips spend le~ time swimming per unit time at sea than those taking diurnal trips thus concluding that penguins do not forage as actively at night (Adams and Wilson 1987, Wilson et al. 1989b). However, Macaroni (Eudyptes cluysolophus) and King Penguins (Aptenodytes patagonicus) at South Georgia (54 °S) expend considerable diving effort at night (Croxall et al. 1988b, Kooyman et al. 1992), although more recent 4studies on a number of species of penguins suggest that nocturnal foraging is characterized by lower capture rates (Wilson et al. 1993, Piitz and Bost 1994). Even though both latitude and season affect daylength and light intensity, making inter-site comparisons more difficult to interpret, it still appears that feeding at night is not commonly employed and that under lower light conditions birds are less successful at capturing prey. While revealing much about penguin foraging and how patterns differ with time of day, the above studies have not examined how patterns change under different light conditions experienced during foraging. This information would provide a more' controlled test of the importance ofvisual orientation during foraging. In addition, all but one previous study (on Gentoo Penguins; Williams and Rothery 1990) lack information about the foraging routine of individual penguins relative to the diellight cycle throughout the chick-provisioning period: Understanding how changes in ambient light intensity regulate a penguin's ability to obtain food is crucial, however, to interpret accurately the relationships between predator performance and prey availability. This study used radio telemetry and food sampling to examine the foraging (i.e., activity patterns, trip frequency, time spent at sea, diet composition and mass) of Chinstrap Penguins, during two breeding seasons, relative to the light-dark cycle and daily light levels. The primary objectiv.es were to: 1) describe Chinstrap foraging patterns and, more specifically, determine the relative importance of nocturnal foraging; 2) test whether reduced light constrains foraging behavior; 3) test whether seasonal changes in light or breeding behavior influence foraging; 4) relate the timing of foraging to the composition and mass offood brought ashore by penguins to determine how changes in diet might also 5reflect adaptations to a dynamic light regime; and 5) examine whether these aspects of their foraging reflect unique conditions characteristic ofa particular season or are persistent between years. Pursuing these objectives at Seal Island, Antarctica provided the opportunity to study a species that is known to forage periodically at night (Bengtson et al. 1993) and that experiences considerable diel periodicity in light during the breeding season (i.e., about 8 loglOlux). 6CHAPTER II MATERIALS AND METHODS Species Description Chinstrap Penguins (4 kg, 71-76 cm) are the smallest of the medium-sized brush- tailed penguins (Genus: Pygoscelis; Clark 1906, Croxall and Furse 1980). During the breeding season, they feed primarily in inshore waters near breeding grounds (White and Conroy 1975, Croxall and Furse 1980) by pursuit-diving to depths typically less than 40 meters and rarely greater than 100 meters (Lishman and Croxall 1983, Bengtson et al. 1993). Most of the world1s population (> 90%) breed on ice-free land along the Antarctic Peninsula and the islands of the Scotia Sea. The population of Chinstrap Penguins within the South Shetland Islands is estimated at 1.6 million individuals, composing more than 90% of penguins breeding there (Shuford and Spear 1988). At Seal Island, an estimated 40000 adult Chinstraps arrive each year in early November to establish nests. Upon hatching one or two chicks in late December, parents begin making daily foraging trips while alternating brooding duties with their mates at the nest. During mid-January chicks are left unattended for the first time (creche phase), and adults, continuing daily feedings, appear to forage independently of one another until chicks fledge at 55-60 days (Lishman 1985a). I I 7 Description of Study Site Fieldwork was conducted at Seal Island, South Shetland Islands, Antarctica (60 0 59'S, 55°23'W; Figure 1), one of many small rocky outcrops in an island archipelago approximately 10 km north ofElephant Island. This study was carried out during the austral summers of 1993 and 1994 at two penguin breeding colonies: North Cove, a colony of approximately 900 nests located on the landward side of a large intertidal pool about 70m from the open sea, and Colony 72, with approximately 400 nests located on a 300m long beach overlooking Beaker Bay (Figure 1, inset). Measurement ofForaging Patterns The presence or absence of penguins at their breeding colony was recorded using radio telemetry on adult Chinstrap Penguins provisioning chicks from 9 January to 14 February in 1993 and from 8 January to 10 February in 1994. Radio transmitters (Advanced Telemetry Systems, model 2) were deployed on one adult from each of80 nests (1993, n = 40; 1994, n = 40) haphazardly selected at North Cove colony. To minimize disturbance, birds were opportunistically captured, after having been relieved by arriving mates. Nest sites were marked with stakes and numbered on a photograph of the study plot to facilitate subsequent observations. A transmitter was attached to the middle of the penguin's back by securing a layer ofouter feathers to the underside of the instrument using a plastic cable tie. A small amount of five-minute epoxy « .5 g) was applied between the layer of feathers attached to the instrument, and the layer beneath it, !I!!)~',l.JllIIJ }I~j 1.lIlj"_I!IIilIII"(I_,n~l~' lh .. ~•• , ,,"" -_ , "'", 0l0/ilII .!lIhln;» '.--:,f-~ 64°8 600S . . 61 0S ./Seal Island Seal Island 6. study colonies -1 0'0 o penguin colonies 0\'0\'00 ~. ,!:>""e~'O('\ ,_ \ IJ '!:>o-v."S' I Beaker 620S -i.~~ Bay Elephant Island 63 0 S -.I I N ~ - t I I 0.5 km I Antarctic Peninsula 65 0 S 600 W 580 W 56Ow 540 W 52Ow 500 W 480 W 460 W Figure 1. Location of Seal Island within the Antarctic Peninsula region and the relative positions of study colonies on Seal Island (inset) 00 I9 to prevent the instrument from pivoting away from the body thereby damaging the feathers or allowing seawater to penetrate. A thin layer of epoxy was also applied to the underside of the instrument and around the cable tie to further insure that the instrument would not fall off prematurely. The instruments were left to fall off during the penguins' annual molt. The transmitters were positioned as far back on the bird as possible to minimize flow disturbances (Bannasch et al. 1994) but not so far back as to make the instrument more conspicuous by having the antenna (29 cm in length) extend beyond the tail feathers. The instruments were wedge-shaped at the anterior end, had a frontal cross- sectional area of 1.4 cm2, weighed 20 g, and were usually positioned posterior to the point of the bird's maximum girth with the antenna lying flat against the animal's back behind the instrument. One device in 1994 stopped transmitting during the bird's first trip to sea and was removed during the subsequent visit. Transmitters were deployed on adults provisioning chicks that were 1 to 2 weeks old (1993, 7-8 January; 1994,6-7 January). The foraging activity of instrumented penguins was measured beginning one day after all 40 penguins were fitted with transmitters in an effort to reduce the effects that handling the birds and disturbing the colony may have had on foraging behavior. Foraging records were collected for 37 and 34 days in 1993 and 1994, respectively, with the end date in both years determined by the beginning of fledging (1993, 14 February; 1994, 10 February). The timing of departures from and arrivals to the island was registered by an automated receiving system that recorded the presence or absence ofeach radio-tagged bird during a ten second interval every fifteen minutes. Because radio tag signals are quenched in seawater and penguins 10 enter or leave the water at North Cove, it was possible to determine accurately the 15- minute interval during which foraging trips were initiated and terminated. Departure and arrival times were used to determine foraging trip durations which are defined as the actual time spent in the water potentially foraging. All arrival and departure times were converted into local apparent time (i.e., 1200 h occurs when solar angle = 90°) for Seal Island to facilitate comparisons with other study sites. By convention, the date on which birds departed was used in the analysis of activity patterns and other mean parameters that were calculated on a daily basis. Visual observations confirmed the status of offspring at each study nest throughout the chick-brooding phase until creching began. In both years all nests had at least one chick until the beginning of creche (1993, 23 January; 1994, 20 January). After chicks creched (i.e., were left unattended), observations of chick-parent interactions were less frequent because adults spent progressively less time in the colony (Lishman 1985a). Nest observations were not continuous, usually lasting 2 to 5 hours each day, thus dates on which both adults were absent from the nest are estimates of the actual timing of creching. Chicks that have creched are intermittently guarded by parents, usually in the early morning and evening (personal observation). Thus, the timing of observations can affect when creching is detected. Some__"Chicks probably creched before it was recorded, but because observations were typically conducted midday, when a minimum number of adults were present, this bias toward slightly later creching dates is believed to be small. I I 11 Criteria to Exclude Non-Provisioning Foragers Because the status ofnests could not be confirmed as frequently during creche, I developed criteria to exclude foraging records from adults that may have stopped feeding chicks sometime after creching. Using data only from birds that were provisioning chicks ensured that the measures of foraging behavior included in each year's sample were not affected by failed breeders which were not subject to the same foraging constraints. Penguin nest failures during creche phase most often occur as the result of predation or starvation (Emslie et al. 1995; Davis and Miller 1990). Brown Skuas (Catharacta lonnbergj) have been observed at North Cove to feed on eggs and smaller chicks, although their success as predators seems to diminish later in the season when chicks are larger (Davis and McCaffrey 1986; personal observation). Giant petrels (Macronectes giganteus), however, are more frequently seen later in the season capturing chicks at the periphery of several colonies at Seal Island (personal observation) and at other breeding sites in the South Shetland Islands (Emslie et al. 1995). North Cove appears to be an exception; Giant Petrels are not common here during the breeding season probably because steep inclines near the inland edge of the colony and a fur seal rookery bordering the seaward edge block access to the periphery of the colony. Starvation, as the result of desertion, was therefore assumed to be the most likely source of mortality among chicks at North Cove. Cases when adults would stop visiting the colony, presumably abandoning their chick(s), were detectable in the foraging activity records prpviding a hasis to exclude data from birds that were less likely to be 12 provisioning chicks. If a penguin did not come ashore on a particular day during the observation period, subsequent foraging trips were excluded from analyses. Those birds which did visit the colony every day were assumed to have continued feeding chick(s) throughout the creche phase. While some birds might have visited the colony daily without feeding chick(s), this behavior seems unlikely given the energetic cost of traveling to and from the colony each day without an obvious benefit to reproductive output. Determination ofFledging Determining the endpoint of the study by observing fledging at North Cove was not possible because thechick(s) of study birds were not uniquely identified and the colony was too large to conduct accurate chick counts. Alternatively, an index offledging chronology was based on daily counts of the number of chicks at a smaller, more isolated colony of approximately 300 nests (Colony 66; Figure 1, inset) during the period after creching had begun. The beginning of fledging was estimated as the date on which the mean of three replicate counts dropped at least 5% between successive days. No studies have specifically addressed the foraging behavior of penguins once they stop feeding chicks prior to fledging. Whether parents continue to make regular trips to sea or change their foraging strategy as they prepare to fast during their annual molt is unknown. Using an estimate of the beginning of fledging as an end date for the study limited the inclusion of foraging records from birds that stopped feeding chicks thus potentially altering their foraging behavior. Growth studies on Chinstrap chicks at Seal Island (unpublished data) and Deception Island (Moreno et al. 1994), however, suggest 13 that parents stop feeding chicks when they are 43-50 days old (i.e., chick mass asymptote), a period as much as a week prior to fledging at Seal Island. Thus, it is still possible that study birds either reduced or terminated the feeding of chicks during the creche period. Food Load Sizes and Diet Composition Diet composition and mass of food brought ashore were determined by extracting stomach contents from penguins at Colony 72 using a lavage technique (Wilson 1984). Birds that had just completed a foraging trip were captured after they reunited with their mates at the nest but prior to feeding their chicks. Birds were held in a vertical, head up, position while 300 ml of tepid water was pumped into their stomachs via a veterinary catheter. The bird was then tipped approximately 30° below horizontal so that its culmen was pointed downward. Slight pressure was then applied to its abdomen after which the bird would typically begin regurgitating its stomach contents into the sample bucket. In 1993, each bird's sample was collected in a single bucket, whereas in 1994, the digested portion (i.e., individual prey in pieces) was collected in a separate bucket from the mostly intact portion of the sample, which was always egested first. During 1993, birds were lavaged until either 1) only clear water was recovered, or 2) they showed signs of distress (i.e., absence of aggression, inability to stand unassisted, or labored breathing), or 3) after the fourth lavage was completed, which was deemed the most disturbance allowable. In 1994, the methods were refined by eliminating the option of stopping when it appeared their stomachs were empty. It was discovered that while occasionally no obvious material 14 would be extracted during the third lavage, food would sometimes reappear during a fourth lavage. While this procedure meant that some birds were lavaged after their stomachs were empty, it provided greater uniformity in the extent oflavaging and reduced bias due to undetectable hard parts that could be present at the bottom ofthe stomach. Considering the change in procedure between years, caution should be exercised when comparing results between years. Five birds in 1993, all arriving in the evening, appeared to have empty stomachs and were lavaged only three times. One bird in 1994 showed signs of distress and was released after the third lavage. Food samples were collected from five different birds every five-day period throughout chick provisioning (1993,6 January - 3 February, n = 35; 1994, 8 January - 12 February, n = 40). Sampling was conducted on birds arriving in the morning (0700-0900 h; 1993, n = 15; 1994, n = 20) and in the evening (1700-1900 h; 1993, n = 20; 1994, n = 20), alternating between five-day periods. Samples were drained, weighed, and sorted into primary prey categories (i.e., krill, fish, squid) and then reweighed to determine percent composition. The mass of each sample determined in 1993 represents the weight after the initial straining of the whole ' sample. In 1994, the total mass of each sample was recorded as the sum ofthe mass of each prey component rather than ~pre-sort weight as in 1993. During the sorting procedure more water drained out of the sample prior to weighing the prey categories. Thus, samples in 1994 could be biased toward a slightly lighter masses compared to 1993. Soft parts, consisting primarily ofkrill and fish, were preserved in a 19% forinalin solution and stored for' future analyses. Hard parts, consisting of fish otoliths and squid t I 15 beaks, were stored separately in isopropyl alcohol and later enumerated and identified to species. In this study, I will present results concerning the total mass and gross composition of the samples in addition to more detailed analyses relating to the fish prey. Estimates ofLight Intensity during Foraging The average light intensity experienced by foraging penguins was estimated by measuring the relative ambient illumination at the top of the island (approximately 120 meters above sea level) using time-depth recorders that were equipped with light sensors (Wildlife Computers, Redmond, WA., USA). While the light conditions measured at Seal Island are not precisely what the penguins would experience on their foraging grounds (3- 26 km offshore; Bengtson et al. 1993) it is believed to be a reasonable approximation. Relative illumination measurements were recorded every 30 seconds during periods that coincided with the radiotelemetry study at North Cove (1993, 18 January - 4 February; 1994, 9 January - 3 March). The overall amplitude of relative illumination measurements was adjusted to correspond with a theoretical range, as determined by known illuminations given solar elevation for the same dates (U.S. Navy 1952), by first arriving at a conversion factor C according to the equation: Max R - Min J( ..._ ----- =C (Relative 11ght uruts . lux i) Max T - Min T where Ma,xR = the maximum relative light measurement taken during the day (1993, 1 February; 1994, 13 January) during the period corresponding to the telemetry study, MinR = the minimum relative light measurement taken at night (1993, 30 January; 1994, 13 16 February) during the same period, MaxT= the maximum theoretical light (assuming no cloud cover) that would occur at solar noon on the same day that Ma0 occurred, and MinT = minimum theoretical light that occurred at solar midnight on the same date as MinR. All relative light data were divided by the C calculated for each year to arrive at corrected lux measurements. This transformation equalized the daily amplitude of the relative and theoretical lux data. The elevation (i.e., relative magnitude) of the two scales was equalized by subtracting the absolute difference between Ma0 and MaxTfrom all corrected lux measurements thus arriving at values termed approximate lux measurements. This approximation of the actual lux values does not take into account the sensitivity threshold of the light sensors at lower light levels (i.e., below one lux). While this would tend to artificially elevate night-time light measurements, the effect would be negligible considering that the difference between the approximate and theoretical illumination for a given sample taken at night would be less than one lux. In addition, as light measurements are averaged over the length of penguins' foraging trips, during which light can vary by up to 8 orders of magnitude, these differences become virtually undetectable. Values of theoretical natural illumination available by time of day were determined by calculating solar elevation according the equation (Wilson 1989a): sin y = sin • sin [) + cos . cos [) . cos t where y = solar elevation C) at given times of day, = geographical latitude C), [) = solar declination C) for a given date, and t = solar angle by time of day C) (e.g., 0 0 = noon and 180 0 = midnight). Solar declination values were determined using the 17 Nautical Almanac (U.S. Navy 1993; 1994). Light measurements taken every 30 seconds were averaged into 5-minute means prior to subsequent analyses. Average light levels experienced during foraging trips were calculated as the mean of five-minute means that occurred during the time the penguin was at sea. Daily means offive-minute mean light measurements were calculated (midnight to midnight) and subtracted from the daily mean of the theoretical light values for each day to provide an index of how light conditions on a specific day compare to the expected conditions given a cloudless day. This was termed the brightness index and was used as a relative comparison measure of light conditions between days. A comparison of overall light conditions between years was not possible due to differences in the sensitivities of the light sensors used each year. However, periodic notes taken on degree of cloud cover did not suggest a substantial difference in light availability between years. 18 CHAPTER III RESULTS Characteristics ofForaging Trips During the study in 1993 and 1994, the frequency distributions ofarrival and departure times of radio-tagged Chinstrap Penguins were bimodal (Figure 2). Upon closer inspection of individual foraging records it became apparent that the periodic nature of the arrivals and departures indicated two general types of foraging: 1) diurnal trips, corresponding to a peak in morning departures and subsequent early evening arrivals, and 2) overnight trips, in which most birds departed in the evening, remained at sea overnight and completed their foraging trip the following morning. Because diurnal and overnight foragers exhibited overlap in the timing of their departures from and arrivals to the island, the midpoint of the foraging trip proved more useful in discriminating the two general strategies and other patterns that might exist. Graphical analysis of the midpoints and durations of trips (1993, Figure 3; 1994, Figure 4) revealed the presence of second category of overnight trips which-,were initiated before noon, spanned the remainder of the day and into the night, and were terminated the following morning. Overnight trips were categorized as either standard overnight trips, corresponding to the dominant pattern of evening departures followed by morning arrivals, or extended overnight trips, an intermittentiJattern in which birds departed in the morning and arrived the following initiated and terminated on the same day, standard overnight trips as those being 19 • • -<>- Departures -- ... - - Arrivals 9h 12h 15h 18h 21h 24h Local apparent time 6h 1993 1994 Oh 3h 16 14 12 10 8 6 4 2 o (/) 0... L.... ....... 0> C 0> cu L.... .E 4- o >- 16 o ~ 14 ::J 0- 12Q) L.... LL 10 8 6 4 2 o mormng. Hence, for purposes of analysis, diurnal trips were defined as those being terminated on the day after the departure date, and extended overnight trips as those overnight trips that were initiated before noon. Figure 2. Frequency distribution of Chinstrap Penguins departing from and arriving at North Cove colony, Seal Island in relation to time of day. 161413121110 tIll I_fa 9 16 B c A ~ ~ ~ ~ ~ I 111~li I ~ ~ I~~~ l~ ji.l ~~' "tlll~ j I~~ * t~ ~ _111~~ 1,1 ~ ~ ~t' ~jl II~ Oh I i I I r I I I r I i I I I I I I 17 18 19 20 21 22 23 ~ ~ ~ ~ ~ ~ ~II ¥~#I 11 ~ Ij~1 ~m tI~~ilj~j~ #~I; , j ill ijU~1 ~lliI ~ Oh I I' I ' I I I iii I I I Oh Oh Oh 15 ~ ~ ;Jlf' I~,.~t~ ~ tl,l' ~~l~ ..if" ~jll ~1111~ .iI~ ¥~j ~l ,~W ~~~U~ ~ jdjl1f'I' ~ ~ ~Wj~II' ~1 ~q' 1tWJ11 d~1' Oh I I iii I t I r I I I I I I I Oh Oh N 12h >.(\j o Oh N ~ 12h o ; ';., 12h (\j o N >. 12h(\j o ~ 12h o .- ~ ~ 12h o --->. (\j 'D '+- o Q) E :.-= ---c (\j Q. U) 'D C (\j C o Q. :2 E Q. .<=: I- tv o 30292423 25 26 27 28 Trip midpoint (time and date: January 1993) Figure 3. Midpoint (.) and span (vertical lines) of each foraging trip, on the scale of a single foraging cycle (y axis) versus the midpoint on a s'cale of days (x axis) for A) 9-15 January, B) 16-22 January, and C) 23-29 January 1993. By definition, trips completed on the day they were initiated (span entirely within day 1) are diurnal trips. Overnight trips extend into the following day (day 2). The clusters of data centered near midday reflect diurnal trips and those centered near midnight reflect standard overnight trips. A third category, termed extended overnight trips and marked by arrows in the figure, comprises trips initiated before noon and terminated the following morning. Oh I I I I I I I I i I I I iii I I 12 13 5 6 ~ ~Ir'" II. ~11i ~~l'~' III1 ~d~I, •• r 11 4 .~ l~il'l~1 3 10 ~ 9 L l*j11 2 ~ iilf!l~'~ 1 8 ~I ~~!t ~~Hlt·'· 31 ~ 7 t ~r " , ,Mll " 30 E o W W w II ~~Idi' 11,~i' I ~ llfit~iI'~' I~ ,.I~~"~~ ~~it' 6 F ~ Oh Oh I i I I i I I i I I Oh N 12h >. ra o Oh N ~ 12h o Oh . Oh ;, 12h ra o [U 12h o Oh ..- N >. 12h ra o [U 12h o .- ....-. >.. ctl u '+- o Q) E :.;::; --c ctl D- lI) U C ctl -c o D- U E D- 'L: f- Oh 1 1 I I 1 I i 13 14 o 1 2 3 4 5 Trip midpoint (time and date: January/February 1993) Figure 3-continued. Midpoint (.) and span (vertical lines) of each foraging trip, on the scale of a single foraging cycle (y axis) versus the midpoint on a scale of days (x axis) for D) 30 January -5 February, E) 6-12 February, and F) 13-14 February 1993. For explanation see first part of figure on previous page. tv Oh IA 'l 12h >- :g Oh Oh Oh 'C 15 16 17 18 19 20 21 22 ~ ~ ~ 11 ~~d ~ll, II ~\Il' - ~l~t 111~j ff~II' ~~~ Idl~I~I~III~ I,~ t~lll'IP' \~ Oh I I iii I t I iii I I l ill N iU' 12h o ..- iU' 12h o 22 23 24 25 26 27 28 29 Trip midpoint (time and date: January 1994) Figure 4. Midpoint (.) and span (vertical lines) of each foraging trip, on the scale of a single foraging cycle (y axis) versus the midpoint on a scale of days (x axis) for A) 8-14 January, B) 15-21 January, and C) 22-28 January 1994. For explanation see Figure 3. N N 543 ~ 2 ~ 1 ~ 3130 ~ 29 E o ~ ~ ~ ~ ~ ~ ~jj I~~jf p~ lItll ~~1 ~~1 ~I 1I I~ I I~ d~ jl111 .~;' ~ ~~ ~Ill ~ Oh I I ' I ' I .' I ' I , I , I ' I Oh ~~r"11 jI tltll~1 -' III ~~" II ~\ ,,' II ~ ttt~" I ~I~I Oh I I ' I ' I ' I ' I ' I 'I I ' Oh Oh N 12h >. ro o Oh N >. 12hro o ~ 12h o --->. ro "0 '5 Q) E :.;::;~ ~ 12h ro ro Cl... 0 (/) u c ro C o Cl... :g E .9- L... f- 5 6 7 8 9 10 11 12 Trip midpoint (time and date: January/February 1994) Figure 4-continued. Midpoint (e) and span (vertical lines) of each foraging trip, on the scale ofa single foraging cycle (y axis) versus the midpoint on a scale of days (x axis) for D) 29 January - 4 February and E) 5-10 February 1994. For explanation see Figure 3. tv w 24 Foraging trips were also divided seasonally into three periods that reflected distinct phases of the Chinstrap Penguins' breeding cycle: 1) chick brooding, which extended from the beginning of the study in each year to the day before the first chick in the study nests creched (1993, 9 - 22 January; 1994,8 - 19 January), 2) transition, defined as the period from the first observed creche to the day 95 % of chicks in the study have creched (1993, 23 January - 1 February; 1994,20 January - 1 February), and 3) creche, the period during which, for all practical purposes, all nests are in the creche phase (in each year only one nest had chicks that creched after transition), ending upon the beginning of fledging (1993, 2 - 14 February; 1994,2 - 10 February). Daily Foraging Patterns In both years, the most common daily foraging pattern was a single diurnal trip and one standard overnight trip was the second most common (Table 1). Extended overnight trips, two diurnal trips, or coupling a diurnal trip with a standard overnight trip were the next most common patterns, totaling about 20% ofthe overall daily activity in each year. Birds not taking foraging trips amounted to about 10% of the daily activity in each year. Other multiple trip patterns comprised less than 1% of the overall activity in both years and were excluded from intra- and--interannual comparisons of activity patterns. The overall activity pattern differed between breeding periods and years. During chick-brooding in 1993, one diurnal trip per day was most common, whereas in 1994 the highest percentage of birds completed one standard overnight trip per day. Extended overnight trips were also more common during chick-brooding in 1994 than in 1993 Table 1. Percent frequency of occurrence of types of daily foraging activity patterns (NT = no trip taken, D = diurnal trip, SN = standard overnight trip, EN = extended overnight trip) for Chinstrap Penguins during the breeding season. Two types of trips separated by a comma indicate that both types were initiated by a bird on a given day. Trip designators preceded by a number (x) indicate that this type of trip occurred (x) number oftimes on a given day. Bird-days represents the total number of days on which individual birds included in the sample were known to takes trips to sea (e.g., 40 birds each foraging o¥er a period of 10 days = 400 bird-days). The distributions of activity patterns were significantly different between all periods. Daily foraging activity patterns (%) Trips-bird day·l Year Breeding period n D SN EN D,SN D,EN 2D 2D,SN 3D NT (bird-days) 1993 Chick-brooding "(9-22 Jan) , 556 47.1 31.8 2.5 7 0 2.5 0 0 9 1.00 Transition (23 Jan - 1 Feb) 390 44.6 27.9 7.2 5.1 0 2.6 0 0.3 12.3 0.96 Creche (2 Feb - 14 Feb) 472 60.6 5.1 6.8 3 0.2 15.5 0 0.8 8.1 1.12 Periods combined 1418 50.9 21.9 5.2 5.1 0.1 6.8 0 0.4 9.6 1.03 1994 Chick-brooding (8-19 Jan) 466 33.7 40.6 12 0.9 0 0.2 0.2 0 12.4 0.89 Transition (20 Jan - 1 Feb) 494 39.1 31 11.1 6.5 0 2 0.2 0 10.1 0.99 Creche (2 Feb - 10 Feb) 337 53.1 5.3 5 6.2 0 24.9 0.6 0.3 4.5 1.28 Periods combined 1297 40.8 27.8 9.9 4.4 0 7.3 0.3 0.1 9.5 1.03 IjJ Statistics: Patterns comprising < 1% ofthe total activity were not included in the comparisons. For each pairwise comparison of activity patterns between periods and years the critical ~X2 value = 11.07, df= 5. For the frequency of activity patterns in 1993 and 1994, X2=42.9 and 275.0 for chick-brooding vs. transition, 489.2 and 9942.1 for chick-brooding vs. creche, and 436.6 and 990.4 for transition vs. creche, respectively; for inter-annual comparisons, X2=237.7, 20.0,41.6 for the three periods, and 102.4 for periods combined, respectively. All P < 0.0001. tv •V) 26 (Table 1). During transition phase in both years, birds took mostly single diurnal trips. However, birds completing any overnight trip (i.e., standard and extended overnight trips combined) during transition in 1994 were still slightly more common than those taking a diurnal trip. The foraging patterns of penguins during creche were similar between the two years with most birds completing one or two diurnal trips per day with standard overnight trips becoming relatively infrequent (Table 1). A chi-square analysis revealed that the frequencies of these daily foraging patterns were significantly different between periods and years (p < 0.001 for each pair-wise comparison; critical value for I:X2 = 11.07 with df= 5; Table 1). Partitioning I:X2 determined which activity patterns contributed most to the differences. In 1993, the change in activity patterns when progressing from chick-brooding to transition was mostly attributed to an increase in the frequency of birds taking one extended overnight trip (X2 = 33.8). However, extended overnight trips were not very common compared to the dominant pattern ofmost birds taking either one diurnal or standard overnight trip per day. The shift in activity patterns between transition and creche was primarily due to a decline in standard overnight trips and an increase in birds taking two diurnal trips (Table 1; X2 = 87.3 and 309.9, respectively). This trend was also evident over the entire study period (X2 = 104.5 and 318.9, respectively, for chick-brooding versus creche). In 1994, the change in activity between chick-brooding and transition was mostly the result of an increase in birds that took two trips per day; either two diurnal or one diurnal trip and one standard overnight trip per day (Table 1; X2 = 181.7 and 75.4, respectively) but again these patterns were not common, amounting to less than 10% of 27 the total activity. Similar to 1993, significant changes that occurred between transition and creche were largely attributed to an increase in the incidence of birds taking two diurnal trips per day (X2 = 880.3) and a decrease in birds completing standard overnight trips (X2 = 70.8). As in the previous year, the increase in the incidence of birds taking two trips per day was largely responsible for the differences that occurred over the entire period in 1994 (Table 1; two diurnal trips, X2 = 9656.2; one diurnal trip & one standard overnight trip, X2 = 114.4). A decrease in birds taking standard overnight trips was also important in differences between chick-brooding and creche (X2 = 102.1). The relative frequency of foraging patterns also changed significantly between 1993 and 1994 within each breeding phase (Table 1). Differences between years in both the chick-brooding and transition periods were primarily attributed to more birds taking extended overnight trips in 1994 (X2 = 167.5 and 19.4, respectively). An increased incidence of birds taking one diurnal trip combined with a standard overnight trip, and birds taking just one diurnal trip per day in 1993 also contributed to differences during chick-brooding between years (X2 = 25.1 and 17.6, respectively). Comparisons of the creche phase showed that differences were mostly attributed to more birds completing two diurnal foraging trips per day in 1994 (X2 = 19.4). Interannual comparisons offoraging patterns for all periods combined indicate that 1993 was characterized by fewer birds taking extended or standard overnight trips (X2 = 53.7 and 20.5, respectively) and more birds taking one diurnal trip per day (X2 = 26.2) relative to 1994. 28 Individual Foraging Behavior In 1993 and 1994, diurnal foraging trips comprised, on average, more than half of a bird's trips to sea with the remainder being either standard or extended overnight trips (Table 2.). No penguins exhibited a strict specialization on one type offoraging trip Table 2. Mean percent trip frequency (S.B.) for Chinstrap Penguins taking diurnal (D), standard overnight (SN), extended overnight (EN) trips. Mean frequencies are the incidence of a type of trip averaged across the means of individual birds during the specified breeding phase (i.e., n = birds). Birds that did not complete trips during each of the breeding phases were not included in the periods combined summary. Types ofForaging Trips (%) 1993 1994 Breeding period n D SN EN n D SN EN Chick-brooding 40 59 38 3 39 40 46 14 (25) (25) (4) (26) (30) (13) Transition 39 58 34 8 38 50 38 12 (28) (28) (9) (24) (25) (8) Creche 39 84 8 8 38 86 9 5 (16) (10) (9) (15) (14) (7) Periods combined 39 68 26 6 38 59 31 10 (15) (14) (5) (14) (16) (7) throughout the study period in either year (Figures 5 and 6, respectively). However, in 1993 during chick-brooding and creche, 8% and 21% of birds, respectively, foraged exclusively during the day (Figure 5). In 1994, there were no diurnal specialists during chick-brooding while 32% ofthe birds foraged solely by day during creche (Figure 6). Although strict diurnal specialization during a breeding period was generally rare, 85% of the birds in 1993 and 76% in 1994 exhibited diurnal foraging on more than half of their ch ick-brooding I I 16 - 40 - 10 - 35 - 25 -9 - 21 - 33 - 22 - 20 -24 - 39 - Co. 14 - Q.l 3'7 - .0 19 - E 36 - :J 4- c 12 - "0 17 - Co. 11 -:.0 3~ =(IJ :J 23 - "0 38 -:~ ~8 - "0 6- c 7- 27 - 3- 18 - 6 - 15 - 29 - 8 - 13 - :~~ -4I I I I 0 20 40 60 80 100 o transition .W"....~ .:.~:-,..'*::-~&::-: :.;.;, :;~:~:~:':~, I I I I I 20 40 60 80 100 creche i~:;:'i~;;;~:;:~~lli~~; ;:!! ':~;",., - ~ -I I I I I I o 20 40 60 80 100 o periods combined ;~:;';I ::.:~~:: ':~9X;:;:' ~ '2 ;~., ';-6:;:;::" x· I I I I I 20 40 60 80 100 Type of foraging trip (%) Figure 5. Incidence of diurnal ( 0 ), standard overnight (. ), and extended overnight ( ~ ) foraging in individual penguin's foraging behavior during each of the specified breeding periods and for periods combined in 1993. Birds are organized in order of increasing overall frequency of diurn~l foraging trips from top to bottom. • tv \D chick-brooding transition creche periods combined .~ :;~:;, :::.::.::.;., ,-:~:;~:., ~m::' .:-:-:-:-~ .~~~~:; ..~~S::';:-"·~I :..s~~;:::.<:::, :;':~::'::'l~n::.;;::.;·' :~~~ _~mmm ..m::;.; ~::~~:'::: ::&Rls«o:' SS~~~>'~~~~' ':::;:;:;;::'~\:«"<:"»~ I I j --r-T o 20 40 60 80 100 0 20 40 60 80 100 :W~,.:., .",.,.~ ,WW:;1 ;S::~~~~oI'., ·:nm..~::«;~::.<:~ .»:,: 8~;';: .,;~... ,;.;:0: .;.;.,;:~::: :..~.....:... :~::.~~~~;:~. ·S::-:"»."""Im~:- :y,:mR..;·: :~mm::.;,;, ;"~;"~~~~::~., ;~i~:9.~~~~i~i:'H:'>':''' :s::.::ms.:R~9:~m;,-;-:s~: 1 j-T--r-l o 20 40 60 80 100 0 20 40 60 80 100 34 - 21 - 5- 19 - 36 - ; 9 - 29 - 27 - 26 - 3- 6 - 17 - '1 -Q; 33- .0 31- E 20- :J 24- c 2- "'0 8- .... 15 - 38- co 4- :J 40- :'Q 30- .2: 14- "'0 32- c 13- 39 - 11 - 18 - 23 - 35 - 28 - 16 - 25 - 37 - 10 - 22 - 12 - Type of foraging trip (%) Figure 6. Incidence of diurnal ( 0 ), standard overnight ( • ), and extended overnight ( ~ ) foraging in individual penguin's foraging behavior during each of the specified breeding periods and for periqds combined in 1994. Birds are organized in order of increasing overall frequency of diurnal foraging trips from top to bottom. w o 31 trips during the study period. A small percentage of birds did forage only at night during certain periods; in 1993, 3% during transition and in 1994, 10% during chick-brooding and 3% during transition. Pattern of Arrivals and Departures The frequency distributions of arrival and departure times, presented in Figure 2, also reflect the changes in the relative importance of types of foraging between periods and years (1993, Figure 7; 1994, Figure 8). Because penguins often initiated two trips in a day, the arrivals and departures of diurnal and standard overnight trips as either first or second trips of the day were considered separately. In general, penguins traveled to and from the island only during daylight hours, very little activity being exhibited at night. During the entire study period in 1993 and 1994 only 1% and 6% of diurnal trips, respectively, were initiated before sunrise and only 2% and 3% terminated after sunset. This lack of activity at night was also evident in departures on standard overnight foraging trips, of which only 1% and 19% occurred after sunset in 1993 and 1994, respectively. Less than 1% of the subsequent arrivals occurred before sunrise in both years. Arrivals and Departures in 1993 Diurnal Trips Patterns of arrivals and departures for penguins taking diurnal foraging trips were similar during chick-brooding and transition in 1993 (Figure 7; la, 2a). During both 32 166 (c) (c) (b) 1 (a) 2 (a) 3 (a) (c) I" • ,I . r-----r--T~r_ 81012141618 0 1 0 a "vi ,~ 8 10 12 14 16 18 20 22 0 2 4 6 o 0000 .1 •• o 0 r-r--""-1I"----r--.--------,----.----.-.-II I~I' 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 IIII I~'I'I 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 L..- -..l:nLlnLJnLJn.LnJ1nU.nLzO'-- ~..L._iLUu1ulLA-.~~__ 50 25 O-L_lULlUl.1.LLIIl.JJ.J.ll.LLclll.J.lJl1I1L.LaLl.JJL-- _ ~HL.. -.l.ln--lLn.lLOillfuLiJe~__-,,-.UL'--"I----"'g-'.-0_uluL.I-I1fL--.._-~_ 25 ~0_ 1 024 Time (local apparent) 50 25 o -L---lLlU.l.JL1LlJ-.ilLlJu...u::u ~HL---_----",-a..11,, -llBu fJu ll.1- ---'CUIiL..li.-.. ii-;;nO__~__(b_)_ 2~ :l."l o 2 4 6 50 2 (a) 25 VI .g- 0 -.L--.JllJllWlUJULl.Lno.UIJlIIlJLJ...Ll...LJ.JJL- ---'- _j ~H nW11JLD 8n" .Ulli ,.. . (b) ~ 2~ ~==, 90 DOn o· 1 I I I. _ _ (e) E ~I III I~Trllll ~ 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 34 periods, more than 70% of birds initiating diurnal foraging trips left between 0300 hand 1000 h and arrived back between 1200 hand 2000 h. During chick-brooding, diurnal trip departures increased at 0400 h and later peaked in higher numbers at 0800 h (Figure 7; la). Low numbers of departures continued until about 2000 h; 10% of departures after the 0800 h mode represented second diurnal trips. As the birds progressed into the transition period the timing of departures remained bimodal, but the relative importance of the modes switched, with the 0400 h mode becoming more pronounced than the peak at 0800 h (Figure 7; 2a). During the creche period, this bi-modal distribution of departures was replaced by a single pronounced peak at 0500 h, during which more than a third of the birds departed (Figure 7; 3a). Birds taking a second diurnal trip contributed to a minor mode at about 1300 h; 47% of departures after 0800 h represented second diurnal trips. Arrivals during creche reflected this bimodal pattern of departures in having at least two peaks: one at 1100-1300 h, composed solely ofbirds completing their first diurnal trip of the day followed by a more prominent mode at around 1500-2000 h, which was composed both ofbirds completing their first (77%) and second (23%) diurnal trips of the day. Standard Overnight Trips Birds departing on standard,gvernight trips also showed similar patterns between chick-brooding and transition with at least three-quarters ofbirds departing between 1500 hand 2000 h (79% and 75%, respectively) and arriving back between 0500 hand 1000 h the following morning (88% and 81%; Figure 7; Ib, 2b). There appears to be no difference in- the timing· of arrivals and departures between birds making their first or 35 second trips of the day during chick-brooding and transition. Although fewer trips were taken during creche, it appears that birds began departing in two waves: one at around 1200-1600 h (n=19 trips), which mostly corresponded to birds' first trip of the day (84%), and another at 1700-2000 h (n=19 trips), which was largely composed ofbirds taking their second trip of the day (59%; Figure 7; 3b). The distribution of arrivals during creche was also similar to earlier periods, although slightly more drawn out, with only 55% arriving between 0500 h and 1000 h. Extended Overnight Trips Departures on extended overnight trips during chick-brooding and transition appeared similar to the later morning departures of diurnal foraging trips (i.e., 0800 h mode) although the frequency of these trips was very low. These trips were completed within the same time period as standard overnight trips although they tended, on average, to return slightly earlier (Figure 7; lc, 2c). During creche, while extended overnight foraging was still infrequent, it appeared that most of the departures coincided with the pronounced 0500 peak in diurnal trip departures (Figure 7; 3c). Arrivals during creche occurred over a broader time span, with birds continuing to arrive up to 5 hours after the latest arrival observed in previous periods. 36 . Arrivals and Departures in 1994 Diurnal Foraging Trips During 1994, the frequency distribution ofdepartures and arrivals for diurnal trips was similar for chick brooding and transition with most birds departing between 0300 h and 1000 h (84% and 66%, respectively) and arriving back between 1200 hand 2000 h (82% and 84%; Figure 8; la, 2a). In contrast to 1993, these periods in 1994 each had a single mode of departures with that of transition (0600 h) being an hour earlier than during chick-brooding (0700 h). Similar to 1993, departures during creche occurred in two modes: a prominent one at 0400 h, during which almost a third of the birds departed, and a later one at 1100-1400 h, which corresponded mostly to birds taking a second diurnal trip on a day (71%; Figure 8; 3a). Standard Overnight Trips As in 1993, most birds departing on standard overnight trips during chick- brooding, transition, and creche in 1994 left between 1500 hand 2000 h (75%,63%, and 68%, respectively) and arrived back between 0500 hand 1000 h (76%, 74%, and 78%; Figure 8; 1b, 2b, 3b). The frequency..ofbirds taking standard overnight trips again decreased through the season. Similar to 1993, there appeared to be no dependence, in timing of these trips, on whether they were initiated as the first or second trip of the day. The majority ofbirds that continued to take standard overnight trips during creche did so only after taking another trip earlier in the day (Figure 8; 3b). 37 Extended Overnight Trips All extended overnight trips during chick-brooding and transition were initiated within the time period of departures for diurnal foraging trips but appeared to be shifted slightly later (Figure 8; Ie, 2c). During chick-brooding and transition, as in 1993, arrivals appeared to conform to those of standard overnight foraging trips. However, extended overnight trips during chick...;brooding appeared to be terminated slightly earlier on average than standard overnight trips (Figure 8; Ie). Similar to 1993, the majority of departures during creche occurred during the primary mode of diurnal departures (0400 h) with arrivals being comparable as well. Duration ofForaging Trips Penguins foraged for significantly different amounts of time within each year depending on the type of trip taken (ANOVA, 1993, F4, 173 = 333.3, P < 0.0001; 1994, F4, 167 = 503.8, P < 0.0001), but trip duration was not different between years (F1, 341 = O. 12, P = 0.73). For this analysis, diurnal and standard overnight trips were again divided depending on whether the trip was the first or second trip of the day. Results of multiple comparisons within each year revealed that all types of trips were of significantly different durations (all P < 0.05) except the two types of standard overnight trips (SN1> SN2; Table 3). In general, first and second diurnal trips were shortest, standard overnight trips were intermediate, and extended overnight trips were longest (Table 3). First diurnal trips were about 40% shorter, on average, than standard overnight trips, which were about 38 two-thirds the duration of extended overnight trips making first diurnal trips less than half the duration of extended overnight trips (all P < 0.05). Second diurnal trips were the shortest being less than half the length offirst diurnal trips (P < 0.05). Because the two basic types of overnight trips were distinguished based on whether the trips were initiated before or after noon, a somewhat arbitrary point, the statistical comparisons of trip duration between the two should be treated with caution. Table 3. Mean foraging trip durations (hours (S.D.); n =birds) for Chinstrap Penguins taking diurnal (Dl), second diurnal (D2), standard overnight (first trip of the day; SN1), standard overnight (second trip of the day; SN2), and extended overnight (EN) foraging trips for all breeding periods combined. 1993 1994 Type n (no. of trips) Trip duration n (no. of trips) Trip duration D1 40 (898) 8.5 (1.2) 39 (685) 9.2 (1.2) Dz 31 (102) 3.9 (1.3) 28 (100) 3.9 (0.95) SN\ 39 (310) 14.8 (1.9) 39 (360) 14.5 (2.7) SN2 29 (73) 13.9 (2.3) 25 (57) 15.4 (2.6) EN 35 (74) 24.2 (4.2) 37 (128) 22.8 (2.3) Effect ofDeparture Time on Duration ofForaging Trips Foraging trip duration decre'lSed with time of departure for each type of foraging trip with all slopes being highly significant (P < 0.0001) in both years (Figure 9). Parameters of the linear regression analysis for 1993 and 1994 are presented in Table 4. A two-way ANCOVA with departure time as the covariate indicated that the effect of 39 40 .. 1993 35 EN ••30 25 .. . • 6 • .. . . .¥20 .\-\ '\: t • "" . .' .1,;'... "1. 6 .15 0 1 ..---... L.- 10 .. :. SN,..c ........ .. .. c 0 5 0:.;::; ro 0L.- :J "'0 0 0- ·C 1994-+-J 0) 35 c EN .0) ...ro 30 .L.- .. 0 LL 25 ..20 6 0 6' 15 0 1 ••SN 2 10 SN, 5 I 0 ___J Oh 4h 8h 12h 16h 20h 24h Departure time (local apparent) Figure 9. Relationship between foraging trip duration and time of departure in breeding Chinstrap Penguins during 1993 (upper) and 1994 (lower); determined for diurnal trips (D1-O, D2-e), standard overnight trips (SN1-Li, SN2-A) and extended overnight trips(EN-.). Refer to Table 4 for explanation of trip abbreviations and regre~sion statistics. 40 departure time on foraging trip duration varied significantly between types of trips (F4 2786 , . = 168.2, P < 0.001) but that this effect did not vary between years (Fl. 2786 = 0.1, P = 0.98). Multiple pair-wise ANCOVAs were used to determine whether the slopes differed between types of trips within each year. These analyses revealed the same trend in both years. The magnitude of the effect ofdeparture time on foraging trip duration was greater for overnight trips than for diurnal trips with one exception: first diurnal trips did not differ from the relatively uncommon standard overnight trips taken as the second trip of the day (Table 4). First and second standard overnight trips were also not different (P > 0.25 in both years). The slope was steepest among extended overnight trips (P < 0.01 for all pair-wise comparisons; Figure 9) indicating that birds arriving after extended overnight trips did so within a narrower window, relative to a broad range of departures, than other Table 4. Linear regression statistics for the relationship between foraging trip duration (y) and departure time (x) for diurnal trips (DI and D2), standard overnight trips (SNI and SNJ, and extended overnight trips (EN). Trip type designations are as described in Table 3. Linear regression statistics 1993 1994 Trip type b b(O) r y/x b b(O) r y/x (minutes) a (minutes) a D1 0.48 11.92 0.26 29 0.57 12.64 0.30 34 Dz 0.26 7.41 0.20 ,~ 16 0.21 6.67 0.10 13 SN\ 0.92 29.75 0.38 55 0.69 25.97 0.34 42 SNz 0.74 27.16 0.21 44 0.67 26.58 0.24 40 EN 1.53 37.29 0.54 92 1.25 32.85 0.49 75 a This statistic provides an index showing how much shorter specified foraging trips are for every 60 minutes trips were delayed (e.g., D 1 trips in 1993 were 29 minutes shorter, on average, for each hour later those trips were initiated). 41 types of trips. In fact, birds that were among the last to depart on extended overnight trips were among the first to arrive the next day. The least influence of departure time on foraging duration was observed in second diurnal trips, during which birds tended to forage for 3-1 0 hours irrespective of the time they departed the island (P = 0.03 compared with first diurnal trips; P < 0.001 for each of the comparisons with overnight trips). Light Regimes for Foraging Penguins Foraging penguins experienced, on average, significantly different light levels depending on the type of trip utilized (F4,663 = 58.3 and F4,1323= 454.4 for 1993 and 1994, respectively, P < 0.0001 for both years). Multiple comparisons (pairwise between each type of trip; all P < 0.05) indicated that in both years birds experienced the highest light levels, on average, during diurnal foraging, with first diurnal trips experiencing significantly more light than second trips only in 1994. Extended overnight trips were generally intermediate being different from all others except second diurnal trips in both years. Light levels experienced during the two types of overnight foraging trips were significantly lower than during other types of trips, but were not different from each other (Figure 10). Birds on first diurnal trips experienced approximately ten times more light (1 loglo lux), on average, than birds on standard overnight trips. There was, however, considerable overlap in the light regimes experienced by individual penguins taking each type of trip. 1993 30 1994 252015 : : .::.1 .' .' 10 .. ' " . . FORAGING TRIP DURATION (HRS) 5 o o 2 2 3 1 4 5 4 Figure 10. Relationship between mean light intensity calculated per trip and foraging trip duration during 1993 (upper) and 1994 (lower); determined for diurnal (0), standard overnight (.), and extended overnight (A) foraging trips. Lines illustrate the theoretical upper (- -) and lower{"') limits oflight, for trips of varying length, calculated for the first day of the study in each year assuming cloudless conditions. 42 3 5 ~ C/) z w r- z r- I (') ---l Z « w 2 43 Effects of Variable Illumination on Duration ofForaging During the period light levels were sampled in 1993, the duration of first diurnal trips were longer during decreased light levels (F1,16 = 14.69, b = -6.19, r = 0.51, P = 0.002; Figure 11). The duration of second diurnal trips also tended to increase with decreasing light levels but the trend was not significant (b = -4.43, r = 0.20, P = 0.22). During the entire study period in 1994, light levels apparently did not have an effect on the duration of either first (b = 0.69, r = 0.007, P = 0.64) or second diurnal trips (b = - 2.08, r = 0.16, P = 0.14). The effect oflight levels on the duration of overnight trips was not assessed due to limitations of the instruments under low light conditions. Partitioning the regression analysis to examine potential effects of light on trip duration within each breeding period was not possible in 1993 because light sampling was mostly confined to the transition period. Examining each period separately in 1994, did not reveal any significant effect of light on trip duration for anyone period. There was, however, a greater tendency toward longer first diurnal trips on darker days during chick- brooding (b = -3.16, r = 0.23, P = 0.16 ) than later in the season (transition, b = -0.55, r = 0.004, P = 0.84; creche, b = 0.67, r = 0.009, P = 0.81). Daily and SeasonatOccurrence ofNocturnal Foraging The daily occurrence of first standard overnight trips and extended overnight trips in relation to the progression of creching and the change in daylength is presented in Figures 12 and 13 for 1993 and 1994, respectively. In both years, there was considerable 12 1993 •11 •• • •10 • • • 9 •~ •'- .c •(J) 8 0. •b ()) 7 •c ()) co • •'-0 6...... -r- 0 co c 1994'- 13:J u •....., •(J)'- 12 • •;;::...... 0 c 11 •0 ro • • •'- .' • :J 10u • •• •c • •co • •Q.l 9 • • •z • • • •8 • • •• •7 •6 5 - 1.2 - 1.0 - 0.8 - 0.6 -0.4 - 02 - 00 Daily brightness index Figure 11. Relationship between daily mean duration of first diurnal foraging trips and brightness index calculated daily for 1993 (upper panel) and 1994 (lower panel). Slope was highly significant in 1993 (P = 0.002) but not in 1994 (P = 0.64). 44 45 variability in the occurrence of standard overnight foraging trips within each period, but overall there was a significant decline in their frequency during transition, 4-7 days after the first chicks creched. The daily frequency of extended overnight trips did not show any obvious seasonal trend in either year (Figures 12 and 13), but rather was positively related to the daily mean duration offirst diurnal trips (Figure 14; P < 0.0001 for both years). Diet ofDiurnal and Overnight Foragers In 1993 and 1994, all Chinstrap Penguins returning from diurnal and overnight trips had predominantly krill (Euphausia species) in their stomachs (Table 5). Fish was found almost exclusively in birds that had been feeding overnight; only one diurnal sample in 1993 had evidence .of fish. While in both years fish was more likely to occur in the stomachs of overnight foragers (~X2 = 118.05 and 269.5, respectively, both P < 0.001), fish was more common in 1994 than 1993 (~X2 = 10.67, P < 0.025) also occurring in significantly greater numbers (comparison of number of otoliths between years: t 1 23 = 1.97, P < 0.035, Table 5). Remnants of squid and amphipods were observed only rarely composing, on average, less than one percent of the total mass of the diet sample. Fish occurring in the diet of overnight foragers were never intact and usually in the form of small pieces of flesh, bones all(i otoliths. The fish prey of overnight foragers, in both years, were primarily lanternfish, family Myctophidae (95%), such as Electrona antarctica, Ii: carlsbergi, Gymnoscopelus nicholsi, and Kreffiichthys anderssoni and less commonly the paralepidid Notolepis coatsi (5%). During lavaging, evidence offish would appear only after a layer of fresher krill had been regurgitated. In 1993, only one $Ajar, QUi".£M4AttJi'uses,s,: kJ lACUU'.."P@AU4J1 t.. ~lZ.b .. tIIU¢!:: .iI ..!!i j[£SkJSIJiilJJJ .. 114);;.J.44#.C4MllIAIin4$..QZxmu4 ..1li\il •.:::;ttJPJA ii"n:;H,WUOiiO@;;\1 ""'" Ol w 0-0 -co co >, u ~ 20% CZ, ~ 16 -o I CB I TR I CR I 21 26 31 December 5 10 15 20 January 25 30 4 9 14 19 February 24 1 6 March Date Figure 12, Daily frequency of standard (-) and extended overnight C···) foraging trips during chick-brooding (CB), transition (TR), and creche (CR) periods relative to the progression of creching (upper panel) and the change in daylength (- -; lower panel) in 1993. .:::.. 0\ Ji$$iJJ4J$ ::;Z4tI1$""1,2" QUiUUJ 1:l2il £J.ttiLlk£ i2 UU au !ki5!U212i£t. 2 asu:w $ k12£iUJiiSJ2 L£l.$f£I.Wa.;UUSMdU 2W&;SJt JWttXJ i,$ UXS ;;,;;:a:: ;;;:M4jWt1;;;;;;W;;S: ;Pt\iM;~ q&ll M""'" 00) 30 .£ .e > U ro 0) 20.eo 2lJl ~u 10c:.c :;:t:u 0 ~lJl ---'- -:J0 18 .e :E OJ I ">. , ro 16 -0 -0 '- 0) ..0 IE 14 :J Z -- ., --',-- , " ------ ---- ---- ----- -...... I CB I TR I CR I 60% -.e 0 'E-o)z 40% >w0-0 -co ro >.U .-Cz 20% O)(j):J ........ O"lJl~a. - 'C 0% ~-ro 0 21 26 31 December 5 10 15 20 January 25 30 Date 4 9 14 19 February 24 1 6 Maroh Figure 13. Daily frequency of standard (-) and extended overnight (..... ) foraging trips during chick-brooding (CB), transition (TR), and creche (CR) periods relative to the progression of creching (upper panel) and the change in daylength (- -; lower panel) in 1994. ~ -l 48 0.4 1993 •0.3 • 0.2 • •en 0.. • • •L. •~ (J) c (J) 0.1 •(1J •L. • • • •0 I'+-~ .c (J) c 0.0 • •L. r-Q)> ; 994 10uQ)uc iQ) 0.4 - • iX •Q) '+- 0 >-0 0.3 •cQ) ::J 0- Q) /L. •'+- >- 0.2 • • •(1J • •0 • • 0.1 • • • •• • •• •• •• • • •0.0 • • • T-----TIv 5 6 7 8 9 10 11 12 13 14 Mean duration of first diurnal (° 1) foraging triPS Figur~ 14. Relation~hip between. daily frequ~ncy ?fextended overnight foraging trips and the daily mean duration of first diurnal foragmg tnps for 1993 (y = -0.02x + 0.Q03x , r = 0.66) and 1994 (y = -O.Olx + 0.003x2, r = 0.64). P < 0.0001 for both years 49 sample from an overnight forager contained parts of fish flesh large enough to be recovered (286 grams), whereas in 1994, 15 samples contained between 2-347 grams of fish flesh. When fish o.ccurred in the digested portion of the sample in 1994 it composed, on average, at least half of the identifiable prey (Table 5). Whether foraging was conducted during the day or overnight had no significant effect on the mass offood brought ashore in either year (Two-way ANOVA, FI,74 = 0.348, P = 0.55). However, year did have an effect on food mass with the mean weight of samples in 1994 being heavier than in 1993 (PI 74 = 18.93, P < 0.001). The increased food mass in 1994 appeared to be the result of larger food loads in diurnal rather than overnight foragers but the trend was not quite significant (year-type offoraging trip interaction, F I 74 = 3.67, P = 0.06). Table 5. Diet mass and composition and frequency of occurrence of fish in the diet of Chinstrap Penguins sampled after returning from diurnal (D) and overnight (ON) foraging trips. Intact and digested portions of the diet samples were examined separately in 1994 only. Mean % composition by weight Intact a No. of otoliths Digested Year Type of n Mean WeIght % intact Krill Fish Squid Krill Fish Squid % occurrence b Forager (g)(SD) (by wt.) of fish Mean Range 1993 D 20 356 (146) na 100 - - na na na 0 ON 15 407 (148) na 96 4 c - na na na 53 I I 5-33 1994 D 20 595 (207) 62 100 t 98 t 5 6 na ON 20 499 (140) 60 96 3 42 43